Audio quality and libsamplerate

Chad.Heuschober at Chad.Heuschober at
Mon Mar 26 17:45:56 CEST 2007


I'm not familiar with Valin's code. Have listening quality tests been 
performed between it and samplerate-best? Is it a library like samplerate 
best or something that's being considered for inclusion in the alsa core? 
Considering that the kernel is GPL I don't think the concern should be 
focused on the 'most open' solution but instead the 'highest quality 
solution' with a minimum openness of GPL/LGPL.


>I think that on the topic of resampling OpenMoko should probably look
>towards the code that Jean-Marc Valin recently wrote for ALSA instead of
>going with libsamplerate. The reason for this is simply that Jean-Marc
>Valin code is licensed under the BSD/LGPL license instead of the GPL and
>thus keeping the OpenMoko platform more open.
>On Mon, 2007-03-26 at 11:08 -0400, Chad.Heuschober at wrote:
> Has the in/exclusion of libsamplerate (aka 'secret rabbit code') in
> alsa been discussed yet? 
> It may seem like a fringe topic but part of the success of devices
> like this is attributed to their ability to do the jobs of many
> devices at once. When it comes to playing music, for the most part,
> portable players are generally very poor performers (with some notable
> exceptions). It made me very happy to see the inclusion of a wolfson
> dac in the device as, traditionally, wolfson dacs have much better
> power handling and higher quality output than other dacs. This is why
> v1 and v2 iPods, which use the wolfson chips, have superior sound
> production to later iterations.
> If libsamplerate is included as part of the core the choice can be
> left to the user to enable some of the other samplers (like
> samplerate-best) and sacrifice the requisite cpu cycles in
> their .asoundrc. There's not a huge cost in KiB's and inclusion of the
> library doesn't dictate that a user MUST use the more expensive
> converter. Personally, I'd like to see it included since it's obvious
> a mixer is necessary and samplerate conversions will occur. But I'm
> also someone who has canalphones that cost more than a high-end iPod
> so it's possible my opinion is skewed. I do firmly believe that,
> assuming the power handling is done well and there isn't voltage
> bleeding (which would be silly considering that this is a mobile
> device and can't afford it) or rampant emi, that this device has the
> potential to reach into the upper echelon of audio production and
> could further distinguish itself by taking that crown from other
> contenders. 
> (Thinking about that inevitable 'comparison' article on a tech
> website). 
> Thoughts? If a 'good idea' who does this get passed to? 
> ~Chad
-------------- next part --------------
An HTML attachment was scrubbed...

More information about the community mailing list